[BalloonEngineer]

Been lurking here for a while, soaking up good info. Wanted a cnc router for years but didn't have the funds. In January attended a "CNC Retreat" hosted at the shop of a friend of mine from the local woodworking guild and Legacy (he has a Legacy Arty 58 and is upgrading to a Maverick). This convinced me to figure out a way. I planned on making a machine out of baltic birch, and had a great idea to make the entire gantry and gantry supports as an integrated torsion box type design. After all, i have a friend with a cnc and he could accurately cut out the pieces I designed.

Heeding the lessons learned on this forum, I spent some time working out the size I wanted, understanding that a wood structure is not stiff enough to be very large. My needs: cutting area at least 24"x36", cutting wood sheet goods and solid wood, maybe occasional plastic. I figured if my parts were within 1/32 (.030) it would probably be better than I was already doing. Do not expect to be able to mill aluminum.

Then I decided to sell my "just for fun" little miata - getting too stiff to get in and out, and it really doesn't fare too well during Colorado winters. Last month listed it on craigslist, and sold it $2100 - just under what i paid for it 10 years ago (it was a 1991 model). Now i had a budget.

While on craigslist, I saw an ad for 8020 extrusions. They had mostly 3030 and 3060 extrusions, the heavy series 15 type. Took a day off work and drove 150 miles and found the mother load. They were used, and covered in dust - they came from some plant that sawed rock with a diamond wire.
Attachment 315318

The provided me with a pallet and I began loading up. Total came out to 385 pounds, charged me $385 at $1 per pound! Inventory:
Qty 2, length 81", 3060.
Qty 1, 64", 3060.
Qty 6, 56", 3030.
Qty 2, 64", 3030
Qty 2, 40", 3030
Qty 4, 31", 3030
Plus a small box of miscellaneous hardware which I'll cover later.

Time to rethink my design.

[BalloonEngineer]

Using the 8020 website, I figured out that the extrusions would have cost me $1881.20, plus shipping (couldn't be that much, only 385 pounds).

Get them home and start cleaning off the dust and some silicone adhesive on some of the parts. A little scraping and a couple of hours with a pressure washer and I'm all set. Find some more hardware as I clean the parts, toss them in the box. The extrusions all look nice and straight, there are marks and a few random screw holes, but all are usable.

If you have never handled this size 8020 beams, they are very impressive and solid feeling. I grabbed the beams somewhat at random, but even as I did so, I started planning my build. I figured the 81" 3060 beams would be my X axis rails, the 64" one would be the gantry. The 56" 3030 pieces would be the cross pieces joining the X rails. The shortest 31" rails would be the legs.

Time to look in my hardware box to see if I can use anything for my build. Ugh, more cleaning. Some time with a wire wheel and see whats there.
Attachment 315354

45 sets at $4.25 at 8020.net, another $191.25 worth. Last thing in the box are a set of leveling feet. Lets see, 7400 lbs. capacity, $31.45 each, should do.


Total spent, $385. Cost to reproduce $2200 (plus shipping). Off to a good start.

Time to start my shopping list for the rest of the build.

[BalloonEngineer]

OOPS! Made a big mistake. My extrusions are not 3060 and 3030 as they were advertised on craigslist. Actually measured them. They are actually 40-8016 and 40-8080. These are slightly larger, heavier, stronger metric extrusions, 80x160mm and 80x80mm. 80mm =3.15", 160mm=6.3". Turns out what I paid $385 for lists for $2,617.57 at 8020.net.

I found that I am getting to follow the advice here and build my second machine first. Instead of a wooden machine with homemade linear rails, I'm going to use commercial linear guides. While I would love to buy profile rails, my budget looks like I can only afford SBR type rails. The $16,000 Legacy Maverick that was at my friend's CNC retreat used that type rail, so I think it will be an acceptable choice for getting started.

Since i now know the sizes of framework that I have, I check out ebay and see what I can find. Decide on a set with 1800mm (~70") and 1500mm (59") rails, SBR20 for the x-axis and SBR16 for y-axis. $300 delivered.

For Z I find a combo with 400mm rails (15") plus 350mm ballscrew. With generous spacing of the guide blocks I hope to have around 5-6 inches of Z travel. Another $115 gone.

For my proposed wood machine I had planned on using one of my routers as a spindle. After seeing how much quieter a true spindle operates and basking in my good fortune with the 8020 extrusions, I decide to splurge on a (cheap, Chinese) spindle. Looks like only the 2.2kw size has collets big enough to handle 1/2 router bits, so decide on one. Water cooled, with a "genuine" not "knockoff" Huanyang VFD, $265. Hey, spending money is fun.

Lastly, going to my friends and listening to everyone there commenting how important after sales support is, I decide to buy the CNCRouterparts package with power supply, 4 420oz-in steppers and Gecko 540. All of these seem to have lots of info here online, and I have seen Ahren helping people by posting here. $705.25. Ouch, spending is not always so fun after all.i

Running total $1770 spent, about $4000 to duplicate. Nothing built yet.

[pippin88]

Have you already bought the rails?

Hiwin 15mm rails are very reasonably priced on aliexpress (try BST automation).

Nice score on the 8020.

[BalloonEngineer]

Yes, the SBR rails were already ordered. I checked out BST Automation. They do have some great prices, but the need to buy multiple sets and their shipping costs put them out of reach for now. Definitely hope to upgrade in future. Need to get it working and making money first.

The place selling the 8020 is actually Motion Constrained Surplus. They also have pretty competitive prices, given that they are close enough for me to pick up. Shipping on long sizes sure gets expensive.

[wizard]

Talk about lucking out with an excellent buy on extrusions. You should end up with a very nice machine on an extremely low budget, that is always delightful. It is always good to hear about these sort of deals. It supports the idea that if you look long enough bargains can be had.

[BalloonEngineer]

Time to start building. Since the rails I bought are only about 70" long, there is no reason to keep my X axis rails at 81". My shop is only 16' x 24', and was pretty packed before I decided to add a CNC router. Realistically, the 24"x36" I had first planned on would fit more easily. But I have all these huge extrusions, and knowing that most always build bigger the second time around, I am going to maximize.

I decide to cut X axis rails to 72". Since I am a woodworker, I have limited metalworking equipment. Basically, looking around, I have a hacksaw and a drill press. That's the main reason I did not follow the normal recommendation to build a steel framed router. Besides, the only good welder I know is my brother, a pipefitter, but he lives 1000 miles away.

I do have a non-ferrous cutting blade on my radial arm saw, because it has a TCG tooth profile and negative hook, which is the preferred configuration on a radial arm saw. Luckily, it is not one of those craftsman pieces of junk, but a solid 60 year old Dewalt. Lets give it a try. The blade is only 8 1/2 in diameter, so I can't cut all the way through in one pass, but I can cut through half way and then flip it over. My saw can actually handle a 12" blade and I have one for wood which could cut in a single pass, but I'll do the two pass cut with the blade designed for metal.

I make sure everything is solidly clamped down and cut. Success! I cut both rails down to 72".
Attachment 315464

Now for the cross pieces. I had initially thought to mount the 56" cross pieces to the bottom of the X axis rails, leaving 50 inches in between. This works out well, since plywood is generally 48" wide and mdf or melamine sheets are 49". But realistically, I would not be able to get the spindle close to the edges of the sheet because of the 6" high X rails would interfere. So I decide to cut the cross rails to 50" and mount them between the X rails, flush on the bottom, 3" below the top.

With a 3/4 mdf sheet on top of the cross rails, the body of the spindle (80mm in diameter) would always be above the top of the X rail with only the collet and bit extending below. With a half inch gap along either edge, I would also be able to surface an mdf spoilboard edge to edge, assuming I use a surfacing bit just a little larger in diameter than my collet.

So if I use 10" spacing of the bearing blocks on my rails for X and Y, my planned cutting area will be about 49" wide x 60" long, and by leaving the ends open, I can handle a full sheet of plywood, tiling the cut if necessary. Hey, thats a larger cutting area than my friends Maverick!

[BalloonEngineer]

Thanks, Wizard. For anyone else interested, the seller was Motion Constrained Surplus in LaPorte, CO, just outside of Fort Collins. I barely made a dent in the 5000 pounds of these extrusions they are selling - $1 per pound, 8020.net's price is over $3 per inch. There were a few pallets worth disassembled, but there was a huge amount still assembled into frames, with some of the larger 40-8016 rails up to 9' long.
I encourage anyone else interested to check it out. They are not selling this online, you must come and get it in person. In my case the savings were well worth the trip. No affiliation other than as a customer.
Attachment 315618

[wizard]

That is a little too far to drive !%$&! I hope people near by take advantage though, at that price it is cheaper than lumber. Further extrusions that she are useful for all sort of things.

[BalloonEngineer]

I agree. I am somewhat tempted to buy some more (those 9 footers would allow expanding the machine I'm building to a full 4'x8') I just don't have the room.

[BalloonEngineer]

Having successfully cut the cross rails to length, I now realize a new challenge. All of the pieces I bought appear to have been factory cut, very smooth and square, and have 4 milled counterbores on each end to use the fasteners I recovered. My cuts are not quite as clean and I do not have access to a mill to machine the pockets into the the newly cut ends. The large X rails do not need them since I am not attaching anything to the cut ends, but my cross rails will need them.

While still under the impression that my extrusions were the inch sized 3060 and 3030, I had ordered some additional fasteners from Tnutz (another $30 spent). Of course I ordered the appropriate (or so I thought) inch sized fasteners. Looking at them after arrival, I see that they have a diameter of 13/16", rather than the 20mm size of the ones I got with the extrusions. I also note that they are chamfered at the end. All of the factory pockets are square bottomed counterbores. The fasteners are designed to fit flush into the recess.

I decide to see if I can use a standard 13/16 drill to make a suitable recess. I happen to have a set of Silver and Deming pattern bits including a 13/16". I have never tried them in metal before, only wood to see if I could get a cleaner hole than spade bits (usually use forstners now).

Now I listed a drill press as one of my few metalworking tools, but my drill press is a pretty pitiful tool. Its not the common Chinese piece of junk, but it I found it at a barn sale and it appears to be homemade, at least there are no manufacturer's marks I have ever found. It has always done what I needed it to do, but this big of a bit in metal will be asking a lot. I decide to do a test in one of the 6" cutoffs from the cross rails.

Attachment 315676

I bolt a drill press vise to the drill press table and give it a go, using the slowest of only 4 speeds It has. Because the holes are centered on the T slot in the extrusion there is a lot of bucking and vibration, but I do get a usable hole. It ain't pretty, but it will do. Now that I know it works I cut the 4 holes on the cut end of all 5 cross rails. Metalworking sucks, there are chips and sharp metal shavings everywhere.
Attachment 315680
Attachment 315678

I envy all those posters on here that own or are able to use a mill to make the metal parts for their router tables, but I much prefer working wood. I will soldier on, improvising as I go.

Running total $1800 spent, about $4030 for anyone foolish enough to try to duplicate, but some progress towards the build phase.

[BalloonEngineer]

Now that I have the cross rails cut to length and drilled for the fasteners its time to start assembling all of these extrusions into a frame. This is a little a puzzle, since you really need to start from the middle out.

First I realize that the ends of the legs need to be tapped to be able to mount the leveling feet. Thats pretty straightforward, as the extrusion profile has holes the correct size in a pattern matching the pattern on the mounting plate. 16 tapped holes, tedious but straightforward. Attach the feet to the bottom of the 31” legs.
Attachment 316120
I want to try to make sure that I assemble the framework as squarely as possible, so I lay the two X-rails on the floor side by side and mark locations for the legs and cross rails across both pieces. Actual positions are not critical, but being even on both sides will keep it square.

Decide to use the two 40" 40-8080 pieces as stretchers between two legs, parallel to the two long X rails. So I lay the X rails on the floor and attach the leg/stretcher assembly sticking straight up. This is easy, until I try to turn them over, first four pieces and it is already getting heavy. The massive X rail makes it top heavy. Once I do turn it over, I realize that it won’t stand up on its own. The leveling feet swivel to accommodate uneven surfaces, so an assembly with only two feet is not stable. Build the other set, attach a board across the 2 with clamps and they stand.
Attachment 316142

This is the first thing I have ever built using aluminum extrusions and I need to learn how to assemble using the T-track and fasteners. Instead of putting the fastener in the counterbore and them attempting to slide the rail into position, is easier to slide the t-nut portion of the fastener into the t-slot of the mating piece and sliding that into rough position. Then put the cross rail into position and slide the fastener into the counterbore and tighten. Get the first cross piece in place, and that holds the whole assembly upright. Then I need to loosen it back up to be able to fit the next pieces in. These fasteners really do create rigid assemblies, and with the first cross rail tightened, I was not able to flex the X rails apart enough to fit the next piece. I think this is a good omen for the rigidity of the finished assembly. Get all five cross rails installed and frame is done. With about 2" in the feet, the 31" leg and the 6" tall X rail, it is taller than I was expecting. May need to cut it down later.
Attachment 316144
Looks very simple compared to many I see here. Leaning my full weight against it and I can barely move it, the movement there is seems to come from flex in my shop floor (just 3/4 flakeboard over 2x4 sleepers). I have seen several builds that use welded steel frames with triangulated bracing and am sure those are more rigid, but I am happy with how stiff mine is. If necessary I can add triangulation using some unistrut that I have. Cutting down the height would also help if necessary, but right now I don't think either will be required.

[Arbo]

Good score on the rails. Wonder how long they will have surplus sitting around. Would considering picking some up to make a base for mine to replace the wood table I made. Am on the western slope so not THAT bad of a drive, once winter is over.

[BalloonEngineer]

They are still running the ad ( Fort Collins Craigslist - search for 8020). The photo in the ad is only a fraction of what they had. More was still assembled into frames. If you go, take some ball end 6mm Hex wrenches - ones you can chuck in a drill or turn with a ratchet will probably speed things up. Also take some tools for scraping off some silicone and digging it out of the hex recesses. Good luck. Grab all the fasteners and hardware you can, buying new gets expensive.

Richard

[BalloonEngineer]

Haven't been keeping up my build log, but I've definitely gotten more done. Let me start catching up.

Time for guide rails to be mounted. Here's where I start to, let us say, depart from the normal.

I mentioned before that I ordered some typical Chinese SBR fully supported round rails, SBR 20 for the X axis, and SBR 16 for the Y and Z axes. I intended to mount them directly to my 40-8016 extrusions that make up my primary X axis and gantry, as my best straightedge shows these extrusions to be fairly straight (Veritas stainless 36" straightedge - not toolroom grade but sufficient for woodworking, claimed to be within .0015 over full length). Along the surfaces I intended to mount the rails on, I could not fit a .002 feeler gauge anywhere.

Of course the 30 mm spacing of the rail mounting hole pattern does not match the 40 mm spacing of my T slots, so I made some alterations to the aluminum mounting flange of the rails. First I removed the ground steel cylindrical rod from the aluminum base so i would not risk scratching it. I marked one end of each piece and kept each set together, so they could be reassembled. Using a 4 1/2" grinder, I made slots in the aluninum bases wide enough so I could mount them to the rails with 5/16 pan head bolts to sliding nuts in the T slots. I cleaned off the metal shavings and reassembled the guide rods to the aluminum bases. On the X rails I made them near the existing holes in case I wanted to use the holes in the future, but it worked so well on the Y axis I just slotted through the existing holes.

On the first X rail, I used my 12" sliding square to space the rail at a constant distance from one edge of one extrusion (centering the rail on the top). I assembled the SBR bearing assemblies to 3/8 pieces of aluminum plate to build my 2 X axis carriages. Across thes I clamped on the gantry rail extrusion to make a temporary gantry and used that to align the second X rail parallel to the first. My temporary gantry travelled very smoothly with light fingertip pressure. I'll count that as a success!

In a similar manner, I mounted the SBR 16 rails to the top and bottom of my gantry extrusion, except here I used the sliding square to mount both rails the same distance back from the "front" face of the gantry, again centering the guide rail on the extrusion.
Attachment 327332
Attachment 327334
Attachment 327344

[BalloonEngineer]

On to the gantry. Here again I am tending to deviate from the norm.
Most DIY units seem to have the gantry cross member mounted between two vertical uprights. I wanted my gantry to be wider than the X rails so that I could have the spindle access all the way to the edge of my work envelope. My Y axis SBR rails are 1500mm (59"). I figured that 10" spacing of the bearing blocks would leave me the 49" travel necessary to accommodate a full sheet of MDF (so I can surface the spoil-board) or melamine if I ever need to, and of course a 48" piece of plywood. Thus the gantry needs to extend past the X axis rails at least 5 inches on each side.

So I am supporting my gantry from the back side. The front, top and bottom need to be clear for travel of the Z carriage. I cut and drilled some pieces of aluminum angle to attach the gantry to some uprights made from another piece of the 80mm 8020 extrusion. This will allow me to adjust the entire gantry up and down should I ever want to carve the top of a log, or assembled box or anything else. The nominal clearance I am planning is 5" between top of spoil board and lowest point in gantry. With the adjustment I could lift entire gantry an additional 6 inches, while the nominal position minimizes moment arms and bending of the uprights.

Attachment 327472

Since the attachment is such that it can be adjusted up and down (the bolts and nutplates are in t-slots), I needed a way to make sure thar both sides were equally spaced (Y axis parallel to spoilboard). My easy solution was to use a spacer directly between the two X carraiges and the Y axis linear rail guide rods. Just used 3 thicknesses of .50 mdf. I did check them with calipers and my makeshift spacers were very consistent.

Attachment 327478
Attachment 327474

[BalloonEngineer]

Y/Z-carriage. Once again I am going to try out some original ideas.

Many, not all, CNC routers have their Y axis rails on the "front" side of the gantry, with a plate as a support for the Y-axis bearing blocks. On the front of this plate are the Z-axis guide rails and then another plate with the mount for the spindle. To me, these designs projects a long way out, with a long cantilever leading to large twisting loads on the gantry, both from the weight of the spindle and guides, and from loads due to cutting.

My goal was to try to reduce this cantilever effect and keep the bit centerline as close to the gantry as practical. I will try to describe my design so that the pictures make sense (or maybe the pictures will help my words make sense). My Y/Z-axis carriage is made entirely of aluminum angle, which I hope will be more resistant to bending and lighter than the typical thick flat plates.

Earlier I described how I mounted my SBR rails to the top and bottom of my gantry extrusion. This provides good separation (over 8 inches center to center), which provides a large moment arm to resist the twisting moment due to the distance from the bit to the gantry. Each pair of bearing blocks is mounted to a piece of angle. These angles were actually cut from a two foot piece of a very heavy aluminum channel (8 inches wide, 2 3/8 high, 7/16 thick) that I found as a cutoff at a metal dealer. The angle pieces are 10 inches long and the bearing blocks mount flush to the ends. The horizontal leg of the angles mount to the Y-axis bearing blocks, the vertical legs are positioned parallel to and just in front of the gantry, both pointing upwards. Joining these two angles are two vertical angles (2 inch x 2 inch, 1/4 thick), again flush to the two outside ends. On the inside faces of these vertical angles are mounted the Z-Axis SBR guide rails (SBR16).

The Z-axis bearing blocks are mounted to yet another pair of angles (3 x 3 x 1/4), joined by the spindle mount and a little piece of plate into a channel. The net result is that the Z-axis guides are mounted to either side of the spindle, not between the spindle and gantry, allowing the spindle to tuck in very close to the front of the gantry extrusion. My spindle centerline is only about 3 inches from the front of the gantry. The Z-axis bearing blocks form about an 8 inch square pattern. Wide separation is good for resisting twisting moments.


This photo is from the rear of the gantry, but shows the heavy angles and the Y-axis bearing blocks.
Attachment 328232

These photos show the assembly from the front, and the vertical travel (a bit over 7 inches) I achieved.
Attachment 328234Attachment 328236

Finally, this shows the overall router at this stage
Attachment 328238

I think this design is fairly original and has some benefits: pretty good strength to weight ratio, compact and easy to make without precision machining (most holes were simply match drilled, if I had access to a mill I would have had a much easier time and would have not had to repeatedly assemble clamp and drill, disassemble etc.). While I described it from the outside in, I actually built it from the inside out, using the actual parts to position everything instead of trying to pre-calculate all the hole patterns. It does somewhat rely on the fact that SBR rails have some angular accommodation, since the legs of extruded angle are never perfectly square. Because the bearing blocks can rotate about the guide rail slightly, this is fine. All three pairs of angle have something mounted to the inside of one leg and the outside of the other, this is not a controlled relationship in any extruded angles. Trying to use this design with profile rails would require far more precision in ensuring these surfaces were perfectly square, but it works well with SBR type rails.

[BalloonEngineer]

Installing the electronics

Well no one has criticized my mechanical assembly yet, I'll take that as a sign I'm not too far afield. I'm pretty sure not everyone will agree with some of my decisions that I will cover today, because again I am departing from some tried and true designs, though in this case I have seen these ideas in commercially made cnc routers.

I'm using rack and pinion for X and Y axis, and as you can see in the last post, I have a ballscrew for Z. That means that all 4 stepper motors are attached to the gantry assembly (2 steppers for X, one on each end). I decided that the easiest way to handle this was to put my electronics box on the gantry as well. This eliminates many long wire runs, the complexity of cable chains and a host of other issues. So I made a box to hold my Ethernet SmoothStepper, Gecko G540 stepper drive, a small 5vdc power supply for the ESS, a relay board, and the larger 48vdc power supply for the steppers. The box is wood and has a clear plexiglas cover so I can see all the LEDs on everything for troubleshooting or just verifying status. The electronics box is mounted to the rear of the gantry. I mounted a filtered fan on the top of the box, blowing in, and have a filtered opening, also on top, for the air to exit. The fan blows directly onto the G540. Since my gantry is a hollow extrusion, I ran much of the wiring from the box through the extrusion.
Attachment 329186

The only cables that connect to the cnc are a 110vac feed to the electronics box, the ethernet cable connecting to the computer and a 220vac feed to the spindle VFD (also mounted to the gantry). I know I could have pulled the 110 from the 220 circuit, but I have them on separate circuit breakers and may put a UPS on the 110v circuit (we get lots of summer thunderstorms and regularly experience brief power outages). These cables drop down from the ceiling with sufficient slack to cover the full range of travel of the cnc. I put switches for the two power feeds as the first point of contact. These are wall type switches but commercial duty 20 amp rated. Thus my on/off for the entire router is on the rear of the gantry. An e-stop switch is mounted to the end of the gantry and wired to the G540 enable (pins 10&12). The e-stop is not connected to any relay to cut power, but does immediately stop the stepper motors.
Attachment 329188Attachment 329190

The computer at the other end of the ethernet cord is an HP all in one computer (display model that was never powered on, $40 on clearance from Walmart) with a clean install of Windows 10 (free upgrade). Since I am building this new in 2016, and I had no previous experience with mach 3, I chose to purchase Mach4 as the control software. The computer is dedicated to running the CNC and is not used for anything else, I have stripped off all other software including basic software installed with windows 10 such as Skype and all the games.

Win10, Mach4, ESS, G540 and the Huanyang VFD are all working very nicely together, with Mach4 starting stopping and controlling speed of the spindle. I have a small Sainsmart 5v dual relay board ($6.99) that is driven from a smoothstepper pin (relay board is powered from the same 5v power supply that supplies the smoothstepper). The Mach4 "Forward" signal closes one of the relays and shorts the VFD FOR(WARD) and DCM contacts which provides the on/off capability, and speed is controlled using the G540 pwm control (did not need to use modbus). The spindle seems to respond from as low as 100 rpm all the way up to its rated 24,000 rpm. The speed displayed on the VFD display usually does not exactly match the commanded speed, but is relatively close over the entire range (+/- 5%).

I could implement reverse using the other relay, but see no point in doing so. Since the ESS also monitors the G540 "fault" (e-stop) signal, the spindle is shut off with an e-stop, and of course Mach4 also sees and displays the e-stop condition.

My 3 mechanical home switches are also mounted on the gantry, all positioned to be triggered mechanically by fixed structure. I do not have limit switches, but do have mechanical stops at each end of physical travel and have implemented soft limits.

The only wiring that does need to move is the stepper wires for Y and Z, the Z limit switch and of course power to the spindle. These are routed through an old piece of 2 1/2" vacuum hose and supported in such a way that they only flex over a rather large radius. The large hose was needed to accommodate the DB9 connectors on my stepper cables (which I purchased from CNC Router parts). My son thinks it looks "ghetto" but it works. I may replace the large hose with some smaller split convoluted tubing, but I used what I had on hand. The spindle cable is shielded 16 awg 4 conductor and I did connect the ground wire internally inside the spindle. I have not had any issues with electrical noise or glitching, although my shop radio does seem to lose signal when the electronics are running.
Attachment 329192

[Arbo]

I'd rather have something that worked that looked ghetto than something that looks pretty that doesn't.